NOTES ON THE VARIATIONS OF RHEGMATODES. 



373 



the other hand, it must not be overlooked that in the Geryonidre 

 we have for the most part the opposite process, namely, the 

 tubes arising from the marginal canal and developing centrally. 

 Then in Figs. 4, 5 and 6, we have conditions which seem to 

 show undoubtedly that such has been the mode of origin and 

 development in Rhegmatodes. It brings before us, therefore, a 

 form of variation at once morphological and physiological, or 

 perhaps better, embryological. 



2. Loops. In Figs. 7 and 8 are shown what has seemed 

 best described as loops. These seem to have resulted from the 

 fusion of two or more canals whose terminal portions have 

 approximated each other and finally become confluent. These 

 were more or less common conditions among these medusae. In 



FIG. 6. Showing varying 

 anastomoses. 



FIG. 7. Showing loop resulting from 

 confluence of distal portions of adjacent 

 canals. 



my earlier paper (op. cit.} I have figured and described very sim- 

 ilar features in Gonionemus, and Agassiz and Woodworth have 

 likewise described similar features in the genus Encopc. 1 I incline 

 to doubt whether any particular significance may be said to 

 attach to such variations as these. It is certainly hard to discern 

 any particular selective value they might have in the betterment 

 of the species. It seems rather to be illustrative of those types 

 of indefinite or fluctuating variations which abound more or less 

 throughout the organic world, but having little or no bearing or 

 significance in natural selection. 



' " Variations in the Genus Eucope," But. Mus. Comp. Zoo!., Vol. XXX. 



